A simple merge sort algorithm with Counting Inversions implementation in C++

MergeSortAndCount.cpp

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

template <typename T>
long long mergeSortAndCountInversions(T* arr, int size)
{
  int m;

	if (size <= 1)
		return 0;

	m = size / 2;

	long long invCountA = 0, invCountB = 0, invCountC = 0;
	invCountA = mergeSortAndCountInversions(arr, m);
	invCountB = mergeSortAndCountInversions(arr + m, size - m);

	T* arrPartA = new T[m];
	T* arrPartB = new T[size - m];

	memcpy(arrPartA, arr, sizeof(T) * m);
	memcpy(arrPartB, arr + m, sizeof(T) * (size - m));

	int i = 0, j = 0, k = 0;

	while (k < size)
	{
		if (arrPartA[i] < arrPartB[j])
		{
			arr[k] = arrPartA[i];
			i++;

			invCountC += j;
		}
		else
		{
			arr[k] = arrPartB[j];
			j++;

			invCountC += 1;
		}

		k++;

		if (i >= m || j >= (size - m))
			break;
	}

	invCountC -= j;

	while (i < m)
	{
		arr[k] = arrPartA[i];
		k++;
		i++;

		invCountC += j;
	}
	while (j < (size - m))
	{
		arr[k] = arrPartB[j];
		k++;
		j++;
	}

	delete []arrPartA;
	delete []arrPartB;

	return (invCountA + invCountB + invCountC);
}

int main()
{
	//int array[] = {4, 2, 1, 3, 5, 8, 0, 43, 10};

	int array[100000];

	for (int i = 0; i < 100000; ++i)
	{
		scanf("%d", &array[i]);
	}

	long long invCount = mergeSortAndCountInversions(array, sizeof(array) / sizeof(array[0]));

	printf("Inversion count: %lld\n", invCount);

	return 0;
}

Though they cancel each other out, line 41 invCountC += 1; and line 50 invCountC -= j; in the code look strange and are redundant.